What I Learned Designing My First 3D-Printed Product

The idea for Totem was simple: a desk object that responds to your presence. Touch it, and it lights up. Come close, and it shifts color. It sits there looking like a sculpture until you interact with it, and then it feels alive.

Simple idea. Humbling execution.

What Totem Is

Totem is a 3D-printed enclosure housing an ESP32-S3 microcontroller, a ring of NeoPixel LEDs, and several capacitive touch pads embedded in the outer shell. The firmware is written in CircuitPython. When you touch different zones on the surface, the lighting responds — different colors, patterns, animations.

The physical form is a roughly totem-pole shape: stacked geometric segments, each a different color, that screw together to house the electronics inside. The goal was something that looked good on a desk whether it was active or off — a design object first, interactive gadget second.

That framing matters, and I’ll come back to it.

Mistake #1: Ignoring Tolerances

I designed the first prototype in Fusion 360, modeled the two halves of the enclosure as a press-fit, printed them, and… they didn’t fit. Not even close. One half was slightly too large for the other.

The problem is that FDM printing adds material — the actual printed dimensions are larger than the designed dimensions, by an amount that depends on your printer, your nozzle diameter, your filament, and your slicer settings. For a basic shape this doesn’t matter much. For mating parts, it’s everything.

The fix is to design an intentional gap — typically 0.15mm to 0.3mm per side for a slip fit, or 0.05mm to 0.1mm for a tight press fit — and then test it by printing small test tokens before committing to a full print. I now keep a small “tolerance test” file that I print on any new printer or filament before doing precision work.

Mistake #2: Wrong Print Orientation

My second design mistake was printing the main tower vertically, which seems intuitive — it’s a tall object, print it tall. What I didn’t think through is that FDM layer lines run perpendicular to the print direction. A vertical print means horizontal layer lines, which means a part that’s weakest exactly where the stress is: at the base, where the tower connects to the weighted bottom.

I had two snap during assembly.

Reprinting at a 45-degree angle solved the structural problem but introduced a new one: surface finish. The angled layers showed up prominently on the smooth exterior panels. For a product that’s supposed to look refined, that wasn’t acceptable.

The real answer was to redesign the part to be printed in sections — shorter, wider segments printed flat, then assembled — which is what the production design now uses. It added assembly steps but gave me both structural integrity and a clean surface.

Multi-Color FDM

Totem’s visual identity depends on distinct color segments. I wanted each stacked section to be a different filament color. There are two ways to do this with an FDM printer:

Manual filament swaps — pause the print at a layer height, swap filament, resume. This is fiddly and introduces a risk of print failure every time you do it, but it works on any single-extruder printer and costs nothing extra.

Multi-filament systems — printers like the Bambu AMS or Prusa MMU can switch filaments automatically. Cleaner results, but introduces purging waste and slows prints significantly.

For prototyping I used manual swaps. For the production design, I’m targeting a multi-filament system so the process is repeatable without babysitting the printer.

Sizing the Electronics Cavity

This one sounds obvious in retrospect: your electronics cavity needs to be bigger than your electronics.

Not by a little. By a meaningful, measured amount.

Capacitors have legs that stick out. Connectors have clips. Wires need to be routed somewhere. The ESP32 Feather board I’m using has a USB-C port that needs to be accessible from the outside — which means a cutout that’s precisely sized, oriented correctly, and in the right position on the part.

My rule of thumb after three iterations: add 0.5mm clearance on all sides of any component cavity, and add 1.5mm around any connector that needs to be plugged/unplugged. Model the actual components in Fusion 360 and don’t eyeball it.

The Pivot That Mattered

About halfway through the prototyping process, I changed how I was thinking about the product. I’d been calling it a “desk toy” — which, fine, accurate — but that framing was affecting my design decisions in subtle ways. A toy is allowed to look toylike. A toy doesn’t need to look refined.

When I started calling it an “interactive desk sculpture,” everything tightened up. The color choices became more deliberate. The seams between segments had to be cleaner. The surface finish had to hold up to scrutiny. The weight had to feel intentional, not incidental.

Naming things matters. The product you ship reflects the framing you started with.

What’s Next for Totem

The current prototype is fully functional — touch response, ambient animations, USB-C charging. What’s left before I’d call it production-ready:

• Finalize the multi-color print workflow on the AMS system
• Source a manufacturer for the weighted bottom section (machined aluminum is the goal)
• Write the production firmware with persistent settings
• Take proper product photos

If you want to follow along as Totem gets finished, subscribe to the newsletter or check the Ephergent store page — I’ll post updates there when it’s ready to ship.

It’s been a frustrating and educational process. That’s usually how the good stuff works.